ASM Metals HandBook Vol. 8 - Mechanical Testing and Evaluation
Подождите немного. Документ загружается.
specimen heating is unwanted electronic interference created by the temperature controllers. Unless an off-the-
shelf engineered testing system is purchased, care is required to shield or filter troublesome signals.
Environmental Chambers
When environments other than still laboratory air are desired, an environmental chamber must be used (Fig.
12). Examples include a vacuum or inert gases (argon, helium) to eliminate effects of oxidation at high
temperatures; high pressure; high-temperature gases, such as oxygen or hydrogen to simulate liquid rocket
engine environments; radiation from radioactive sources for the study of nuclear reactor components; cryogenic
environments for the study of alloys used in super-cold applications; or even corrosive environments that can
degrade the fatigue properties of a material. Each environment poses its own set of concerns for the protection
of grips, extensometry, and load cells. If, for example, a vacuum or pressurized chamber is required, any
parasitic loads transmitted through the structural components required to maintain the environment (e.g., seals,
baffles, or bellows) must be eliminated from the load cell signal.
Fig. 12 Environmental chamber for fatigue testing
Careful attention must be given to gaining access to the specimen and grips for installation and removal of
specimens, and the attachment and support of extensometry. Viewing ports are also highly desirable. For sealed
systems, insulated electrical feed-through connectors are required for transducer signals (e.g., strain, pressure,
and temperature).
Representation of Material and Specimen Configurations
A specimen is a representative sample of an alloy and is used to determine and compare the basic fatigue
characteristics of a material. In addition to this primary use, specimens can be designed to study the effects on
fatigue resistance of several important factors. These factors can be categorized as:
• Geometric features, such as stress concentrations including intentional notches, holes, fillet radii, and
attachment configurations; unintentional voids or flaws; and damage due to handling or service use
• Bulk properties, such as those influenced by degree and orientation of cold working, temperature of
testing, thermal exposure, heat treatment, and radiation
• Surface-related effects, including mechanical surface finish, shot peening, burnishing, laser shock
treatment, residual stresses, environmental interactions, fretting, galling, and coatings
• Loading-related factors, such as mean stress, multiaxiality, cumulative damage, temperature of testing,
thermal cycling, and creep-fatigue interaction that might occur at low testing frequency and hold times
Nomenclature for a typical fatigue specimen is given in Fig. 13. A test specimen has three sections: the test
section, a transition zone, and the two grip ends. The test section is where the quoted stresses, strains,
temperature, and environment are measured or controlled and where fatigue cracking and failure are designed
to occur. The majority of the specimen deformation occurs in this section. The grip ends are designed to
transfer load from the test machine grips to the test section and may be identical at either end, particularly for
axial fatigue tests. The transition from the grip ends to the test area is designed with large, smoothly blended
radii to minimize stress concentration in the transition, which could otherwise initiate undesired fatigue
cracking.
Fig. 13 Nomenclature for a typical fatigue specimen. Ends of round specimens may have smooth shanks,
button heads, or threads. Smooth shanks should be long enough to accommodate some type of wedge
grip. Rectangular specimens are generally made with smooth shanks, but may be shouldered to contain a
hole for a pin bearing.
The design and type of specimen used depends on the fatigue testing machine and the objective of the fatigue
study. The test section in the specimen is reduced in cross section to increase the stresses and strains there and,
hence, avoid failure in the transition region and in the grip ends. The test section should be proportioned to
properly test the material, accounting for adequate sampling of specific microstructural features (e.g., defects
and grain size), yet still be small enough that the load capacity of the load frame is not exceeded.
The location from which a test specimen is taken from the initial product form is important because the manner
in which a material is processed influences the uniformity of microstructure along the length of the product as
well as through its thickness. For example, the properties of metal cut from castings are influenced by the rate
of cooling and by shrinkage stresses at changes in section. Generally, specimens taken from the surface of
castings are stronger. Many ASTM standards, such as E 8 and B 557, provide guidance in the selection of test
specimen orientation relative to the rolling direction of the plate or the major forming axes of other types of
products, and in the selection of test specimen location relative to the surface of the product.
Orientation is also important to standardize test results relative to the directionality of properties that often
develops in the microstructure of materials during processing. Some causes of directionality include the
fibering of inclusions in steels, the formation of crystallographic textures in most metals and alloys, and
preferred growth directions in directionally solidified and single-crystal materials.
Specimen Machining and Surface Preparation
Because fatigue crack initiation typically is surface dependent, proper machining and surface preparation of test
specimens are critical. Unless care is taken, scatter caused by variable surface conditions will overwhelm the
inherent scatter of the material being studied.
Because a primary aim of fatigue testing is comparison of materials, uniform preparation procedures must be
established. Machining operations must not alter the surface structure of the metal; thus, heat generation, heavy
cutting, and severe grinding are prohibited. Final machining should be parallel to the direction of applied stress.
Transition fillets must be blended into the test area without steps or undercutting. Surface polishing using
metallographic techniques is preferred for smooth specimens, where machining marks are removed by a
sequence of grinding steps. Example procedures for machining of specimens are given in ASTM E 466 and Ref
19.
The final polishing is not a buffing operation, but a cutting operation that uses lapping compounds or aluminum
oxide powder in a liquid medium to remove grinding scratches. For flat sheet or plate specimens, edges should
be slightly rounded and ground to eliminate nicks, dents, cuts, and sharp edges, which can lead to premature
crack initiation. The relationship between surface characteristics and fatigue properties is discussed later in this
article.
Test specimens received from a machine shop are expected to meet size specifications provided to the shop. To
ensure dimensional accuracy, however, each test specimen should be measured prior to testing. Gage length,
fillet radius, and cross-sectional dimensions are easily measured. Cylindrical test specimens should be
measured for concentricity. Maintaining acceptable concentricity is extremely important in minimizing
unintended bending stresses.
In general, flat specimens are difficult to manufacture without twisting or bending; especially if they are taken
from rolled products. The potential warping of thin specimens and their inadequate section modulus lead to
ready elastic buckling under compressive loads. Therefore, these types of specimens are usually only tested in
tension. Lateral antibuckling guides can be used to prevent buckling of compressively loaded thin flat
specimens (Ref 20). However, the buckling guides introduce other problems (surface contact and small
amounts of rubbing action at points of contact) that can affect the fatigue properties.
The product form from which the specimen is taken often influences the geometry of a test specimen.
Obviously, only flat specimens can be obtained from sheet products. Test specimens taken from thick plates or
bar stock may be either round or flat. Occasionally, subscale specimens must be employed if the product form
is too small for standard specimen dimensions.
Cylindrical Specimens. Three types of specimens with circular cross sections are commonly used:
• Specimens with tangentially blending fillets between the test section and the grip ends (Fig. 14a and b)
• Specimens with a continuous radius between the grip ends with the minimum diameter at the center
(Fig. 14b and d and 15a). These are referred to as hourglass specimens and are commonly used when
strain ranges are larger than 2%.
• Specimens for use in cantilever-beam loading with tapered diameters proportioned to produce nominally
constant stress along the test section
Fig. 14 Typical fatigue test specimens. (a) Torsional specimen. (b) Rotating-beam specimen. (c) Plate
specimen for cantilever reverse bending. (d) Axial loading specimen
The design of the grip ends depends on the machine design and the gripping devices used. Round specimens for
axial fatigue machines using grip inserts like those shown in Fig. 10 may be threaded, button-head, or smooth-
shank types. For rotating-beam machines, short tapered grip ends with internal threads are used, and the
specimen is pulled into the grip by a draw bar. A long, smooth shank end is used on machines with lathe-type
collets.
Flat Sheet and Plate Specimens. Generally, flat specimens for either axial or bending fatigue tests are reduced
in width in the test section but may have small thickness reductions as well. The most commonly used types
include:
• Specimens with tangentially blending fillets between the test section and the grip ends (Fig. 15c). This
specimen has a straight gage section and is used in both axial and bending fatigue.
• Specimens with a continuous radius between the grip ends, giving a flat hourglass design (Fig. 15d).
These are also used in both axial and bending fatigue test.
• Specimens for use in cantilever reverse-bending tests with tapered widths (Fig. 14c)
Flat specimens generally are clamped in flat wedge-type grips or may be held with a stiff bolted clamp/joint
friction grip for reversed axial loading. Pin loading can be used when only tensile loads are encountered. When
pin loading is utilized, the holes drilled in the grip end must be designed to avoid shear or bearing failures at the
holes, tensile failure between the holes at maximum load, and fatigue cracking at the holes in the grip end. In
axial fatigue testing of flat sheet specimens, the test length and cross section must be designed to prevent
premature buckling of the specimen.
Fig. 15 Typical round and flat fatigue test-specimen configurations. (a) Hourglass specimen with
continuous radius between grip ends. (b) Round specimen with tangentially blended fillets between test
section and grip ends. (c) Flat specimen with tangentially blended fillets between test section and grip
ends. (d) Flat specimen with continuous radius between the grip ends
Alignment Considerations
The keys to achieving accurate fatigue data include ensuring that applied loading is aligned with the specimen
axis. This is particularly important in axially loaded specimens. Poor and nonreproducible alignment produces
bending strains that reduce the fatigue life and increase the scatter in the data. In fact, it has been calculated that
the largest contributing factor to scatter in LCF data is due to bending (Ref 21). Bending strains arise when
there is a misalignment somewhere in the load train. The axiality and concentricity of the actuator, grips, test
specimen, and load cell should be verified and corrected if there is a problem. Coarse adjustments can be made
by removing the preload and loosening the load train. These components can then be shifted and/or shimmed to
gain proper alignment. Once the load train has been retightened, the small bending strains, which always exist,
must be measured and minimized. This is typically done by using a strain-gaged trial test specimen, which
allows bending strains to be calculated as a function of position along the test specimen, using the same
material and test set-up as used in the actual test program. This process can be somewhat involved, and ASTM
E 1020 describes it, along with many articles (Ref 22, 23, 24, and 25).
During axial loading, the bending strains should be kept below a specified amount, as described in various
fatigue testing standards. For example, ASTM E 606 recommends that the maximum bending strain should not
exceed 5% of the minimum axial strain range used during the test. If the bending strains exceed these amounts,
the test rig must be further aligned. In past years, this was done by a trial and error method similar to the coarse
adjustment procedure described above. This process could take days of effort to achieve moderate bending
results. Recently, a new device has been developed (Fig. 16), which fits into one end of the load train (Fig. 8)
and allows adjustment of both angular and concentric components of bending while the load train is under
preload. This reduces the time needed for achieving proper alignment to within a few hours.
Fig. 16 Alignment fixture for minimizing bending strains in axial fatigue testing
The amount of bending strain in the specimen is affected by gripping methods and specimen design. Better
alignment can be achieved using collet, wedge, and button-head grips. Threaded specimens generally give
poorer alignment with equally poor reproducibility. Likewise, wear and oxidation of the grips can lead to poor
alignment.
Graphic Recorders
In addition to software-based digital recording methods, analog recording devices are also commonly employed
for materials testing applications. Strip-chart recorders, enabling time-based paper chart records of various
control and response variables, are used. Generally, the strip chart record is used to provide a means to diagnose
why a test went off-line, if it does so prior to failure, and to record the gross response variable behavior (e.g.,
load response) at the point of incipient failure under strain control conditions. Most pen-based strip chart
recorders are useful for lower-frequency testing applications (e.g., <1 Hz), because their limited frequency
response precludes use to higher frequencies. XY recorders are also commonly employed to record material
stress-strain (load-displacement) response or hysteresis loop response in fatigue tests. XY and strip-chart
recorders often feature a selection of preset gain ranges (e.g., 0.1 V, 1 V, 10 V or 1 V, 2 V, 5 V, 10 V) for use in
plotting signals from the test system. For materials testing applications, it is most desirable that the preset gain
controls offer ranges that approximately double the value of the previous setting with each subsequent setting
(e.g., 1 V, 2 V, 5 V, 10 V). Analog recorders of these types are often used in conjunction with digital controllers
and associated software due to the additional flexibility offered.
References cited in this section
2. Strain Gage Based Transducers, Their Design and Construction, Measurements Group, Inc., Raleigh,
NC, 1988
3. M.F. Day and G.F. Harrison, Design and Calibration of Extensometers and Transducers, Measurement
of High Temperature Mechanical Properties, M.S. Loveday et al., Ed., National Physical Laboratory,
London, 1982, p 225–240
4. G.B. Thomas, Axial Extensometry for Ridged Specimens, Techniques for High Temperature Fatigue
Testing, G. Sumner and V.B. Livesey, Ed., Elsevier Applied Science, 1985, p 45–56
5. J.W. Dally and W.F. Riley, Experimental Stress Analysis, McGraw-Hill, 1978, p 179–179
6. R.L. Hannah and S.E. Reed, Ed., Strain Gage Users' Handbook, Elsevier Applied Science, New York,
1992
7. J.-F. Lei and W.D. Williams, PdCr Based High Temperature Static Strain Gage, AIAA 2nd Int. National
Aerospace Planes Conf. Publication, AIAA-90-5236, 1990
8. J.-F. Lei, M.G. Castelli, D. Androjna, C. Blue, R. Blue, and R.Y. Lin, Comparison Testings Between
Two High-Temperature Strain Measurement Systems, Exp. Mech., Vol 36 (No. 4), 1996, p 430–435
9. E.G. Ellison, Thermal-Mechanical Strain Cycling and Testing at Higher Temperatures, Measurement of
High Temperature Mechanical Properties, M.S. Loveday et al., Ed., National Physical Laboratory,
London, 1982, p 204–224
10. E.G. Ellison and R.D. Lohr, The Extensometer-Specimen Interface, Techniques for High Temperature
Fatigue Testing, G. Sumner and V.B. Livesey, Ed., Elsevier Applied Science, 1985, p 1–28
11. Creep, Stress-Rupture, and Stress-Relaxation Testing, Mechanical Testing, Vol 8, Metals Handbook,
ASM International, 1985, p 313
12. J.Z. Gyekenyesi and P.A. Bartolotta, “An Evaluation of Strain Measuring Devices for Ceramic
Composites,” NASA TM 105337, National Aeronautics and Space Administration, Nov 1991
13. R. Hales and D.J. Walters, Measurement of Strain in High Temperature Fatigue, Measurement of High
Temperature Mechanical Properties, M.S. Loveday et al., Ed., National Physical Laboratory, London,
1982, p 241–254
14. ASTM STP 465, Manual of Low Cycle Fatigue Testing, R.M. Wetzel and L.F. Coffin, Ed., American
Society for Testing and Materials, 1969
15. M.G. Castelli and J.R. Ellis, “Improved Techniques for Thermomechanical Testing in Support of
Deformation Modeling,” ASTM STP 1186, Thermomechanical Fatigue Behavior of Materials, H.
Sehitoglu, Ed., American Society for Testing and Materials, 1993, p 195–211
16. S.S. Manson, G.R. Halford, and M.H. Hirschberg, Creep-Fatigue Analysis by Strain-Range Partitioning,
Design for Elevated Temperature Environment, American Society of Mechanical Engineers, 1971, p
12–28
17. K.D. Sheffler and G.S. Doble, “Influence of Creep Damage on the Low-Cycle Thermal-Mechanical
Fatigue Behavior of Two Tantalum Base Alloys,” NASA CR-121001, National Aeronautics and Space
Administration, 1972
18. M.H. Hirschberg, D.A. Spera, and S.J. Klima, “Cyclic Creep and Fatigue of TD-NiCr (Thoria-
Dispersion-Strengthened Nickel-Chromium), TD-Ni, and NiCr Sheet at 1200 °C,” NASA TN D-6649,
National Aeronautics and Space Administration, 1972
19. P.R. Breakwell, J.R. Boxall and G.A. Webster, Specimen Manufacture, Measurement of High
Temperature Mechanical Properties, M.S. Loveday et al., Ed., National Physical Laboratory, London,
1982, p 322–340
20. E.T. Camponeschi, Jr., “Compression of Composite Materials: A Review,” ASTM STP 1110,
Composite Materials: Fatigue and Fracture (Third Volume), T.K. O'Brien, Ed., American Society for
Testing and Materials, 1991, p 550–578
21. F.A. Kandil and B.F. Dyson, Influence of Load Misalignment During Fatigue Testing I and II, Fatigue
and Fracture of Engineering Materials and Structures, Vol 16 (No. 5), 1993, p 509–537
22. J. Bressers, Axiality of Loading, Measurement of High Temperature Mechanical Properties, M.S.
Loveday et al., Ed., National Physical Laboratory, London, 1982, p 278–295
23. J. Bressers, Ed., A Code of Practice for the Measurement of Misalignment Induced Bending in
Uniaxially Loaded Tension-Compression Test Pieces (Institute for Advanced Materials, Joint Research
Centre, Netherlands, and the High Temperature Mechanical Testing Committee), the Directorate-
General of the European Commission, Luxembourg, 1995, 52 pages
24. F.A. Kandil, Measurement of Bending in Uniaxial Low Cycle Fatigue Testing, National Physical
Laboratory, Teddington, England, 1998
25. A.K. Schmieder, “Measuring the Apparatus Contribution to Bending in Tension Specimens,” ASTM
STP 488, Elevated Temperature Testing Problem Areas, American Society for Testing and Materials,
1971, p 15–42
Fatigue, Creep Fatigue, and Thermomechanical Fatigue Life Testing
Gary R. Halford and Bradley A. Lerch, Glenn Research Center at Lewis Field, National Aeronautics and Space Administration; Michael
A. McGaw, McGaw Technology, Inc.
Electronic Test Controls
Electronic test controls (controllers) are used to adjust and maintain the desired control parameter(s) for a given
fatigue test. Controllers also provide test termination capabilities for many criteria, for example, failure, load
drop-off, deflection, or extension limit excellence. Modern fatigue testing is generally performed using closed-
loop servocontrollers, wherein the controlled parameter is continually sensed and compared to the desired
command, and the result of this comparison, the error signal, is used to drive the actuator (which may be
hydraulic or electromechanical in nature). Sensors for measuring the relevant mechanical quantities (e.g., load
cells and extensometers, as discussed elsewhere in this article) provide the feedback, and the control mode of
the test is defined by the sensor being used for feedback control. Control is typically effected through a PID
strategy wherein the error signal is amplified by gain terms that can be independent of time (proportional gain),
as well as dependent on time (integral and derivative gain). A typical closed-loop control system is shown in
Fig. 17.
Fig. 17 Typical closed-loop servocontroller system for fatigue testing
A functional fatigue testing machine requires the ability to flexibly define test needs with regard to command
waveform generation and data acquisition. A variety of technologies are available to accomplish closed loop
control of materials testing systems, in performing standard materials tests, and for the development of custom
testing applications. This section discusses these technologies and particularly focuses on the state of the art of
software tools for materials testing.
Load Frames: Analog and Digital Controls
Analog controllers are the most commonly used test controllers in fatigue testing laboratories today. Analog
controllers have been brought to a high level of refinement, and the latest examples exhibit relatively low noise
and reasonably wide-frequency bandwidth. Analog controllers typically provide numerous inputs and outputs
that can be used to flexibly adapt to nearly any testing requirement. In addition, materials testing applications
software that is designed for analog controllers can be used broadly on many different controller models. The
disadvantages of analog controls include the fact that many models do not provide the ability to switch control
modes while the test system is energized (under hydraulic pressure), and that calibration of the sensors and the
test controller is typically done through trim pot adjustments, a process that can be time consuming.
Digital controllers have been available for the past several years, and the usability of these controllers continues
to improve with each new generation. The fundamental difference between an analog controller and a digital
controller centers on the summing junction (Fig. 17). A digital controller closes the control loop using a
microprocessor instead of continuous-signal analog amplifiers. In a digital controller, the loop closure rate (the
rate at which the microprocessor must update the control loop) is directly related to the maximum test
frequency performance (i.e., bandwidth) that can be attained. The loop closure rates required for servohydraulic
systems demand very-high-performance digital systems. One of the factors driving the development of digital
controls is that control mode switching is simpler, as the loop is digital, and many of the problems relating to
offset error in analog systems are not present in their digital counterparts. In addition, from a manufacturer's
viewpoint, the cost of digital technology continues to drop (for a given level of performance), and the
manufacturing of digitally based products is inherently simpler, owing to fewer required circuit trim
adjustments. The chief disadvantages of digital controllers are that changes to the controller organization cannot
be done except through software modification of the controller, effectively isolating the user from making any
change, and digital controllers generally require software graphical user interfaces. Most digital controller
operator interfaces can be labyrinthine to navigate and use, complicating their use and increasing the probability
of user mistakes.
Comparison of Analog and Digital Controllers. Both analog and digital control technologies are available for
use in performing materials tests, and it is not possible to make definitive statements regarding the superiority
of either technology for materials testing needs. The decision as to which controller technology to use may be
more prosaic as well: life cycle costs and training costs, among others, must be carefully considered before a
choice is rendered. Because analog controllers provide a nearly universal interface for admitting external
program command signals, as well as provide high-level conditioned transducer signals for data acquisition,
they are easily interfaced to computers for test control purposes. The software created can be generally applied
to testing needs regardless of the specific manufacturer's model of the analog servocontroller. Digital
controllers, on the other hand, feature software that is based on the “command set” for the specific controller.
Thus, materials testing needs that cannot be realized from off-the-shelf software must be custom developed, and
any such software is uniquely tied to that specific digital controller model. Finally, analog controllers have a
demonstrated longevity, whereas the digital controller will likely have a much more rapid obsolescence, owing
to the rapid evolution of digital components.
Furnace Controls
The control of furnaces used in materials testing is largely accomplished using closed-loop temperature
controls. Because thermal processes inherently vary slowly (therefore the control loop update rate requirements
are relatively modest), and because the costs of microprocessor technology continue to drop, nearly all
temperature controllers available today are digitally based. Care must be exercised when selecting a given
temperature controller to ensure that it is compatible with the control input requirements of the heating system
that is being used. If thermomechanical tests are contemplated, or other nonisothermal temperature
requirements are being contemplated, the temperature controller must have facilities for varying the
temperature set point, either by means of an external command signal or with a controller command set. Of the
two, facilities for accepting an external command signal are preferred, because this provides the greatest
flexibility and simplicity in the test apparatus.
Test Program Development and Software
Software systems for performing materials tests have been in use for approximately 25 years (Ref 26). During
this period, computer technology has changed dramatically, but, ironically, testing needs have not. Scientists
and engineers have a continuing need for flexible and powerful tools to design and conduct materials tests, be
they standard tests or unique experiments representative of research and development efforts. Software systems
developed to satisfy testing needs have typically been developed along three lines: single application software
created uniquely to meet a specific testing requirement (e.g., ASTM E 606), general-purpose testing software
designed to provide a flexible set of tools that can be used to implement a broad spectrum of testing
requirements, and lastly, custom-written test application software.
Single-Purpose Software. Software for performing standardized tests is widely available from several vendors,
operating for both analog and digital servocontrollers. Materials testing engineers charged with conducting
standardized materials tests are strongly encouraged to review commercially available offerings before
considering custom-application software development. Most specific test requirements (e.g., high-cycle fatigue,
low-cycle fatigue, and fatigue crack growth) are well represented by commercially available application
software. While the cost of commercial materials testing software is significant, custom-application
development is not trivial and is time consuming.
General-Purpose Software. Several examples of general-purpose testing software are readily available from a
variety of vendors. Generally, these applications operate on a PC interfaced (in varying degrees of complexity)
to a servocontroller, either digital (Ref 27, 28, and 29) or analog and/or digital (Ref 30, 31, 32, 33, and 34).
These tools permit the construction of reasonably complicated test sequences and provide for data acquisition